You are here

Mauro Cettolo

  • SKYPE: mcettolo
  • Phone: +39 0461314551
  • FBK Povo
Short bio

I have been a researcher at FBK/ITC-irst since 1992. My main research interest has been statistical machine translation since 2004. Previous research activities regarded the efficient organization of the search space for large vocabulary, continuous speech recognition, the automatic segmentation, classification and clustering of audio broadcast news, the automatic transcription of spontaneous lecture speech. I co-authored more than 100 scientific publications and have served as a reviewer international journals, conferences and workshops. I was involved in the organization of international scientific events, like EAMT, LaDaKT, AISV, evaluation campaigns of IWSLT. I taught computer science modules as consulting professor at the universities of Trento and Verona.

Research interests
statistical machine translation; efficient organization of the search space for speech recognition; automatic segmentation/classification/clustering of speech

  1. Fabio Brugnara; Mauro Cettolo; Marcello Federico; Diego Giuliani,
    Advances in Automatic Transcription of Italian Broadcast News,
    ICSLP 2000,
    , pp. 660-
    , (ICSLP 2000,
    Beijing, China,
    16/10/2000 - 20/10/2000)
  2. Mauro Cettolo; Marcello Federico,
    Model Selection Criteria for Acoustic Segmentation,
    ISCA ITRW ASR2000,
    , pp. 221-
    , (ISCA ITRW ASR2000,
    Paris, France,
    18/09/2000 - 20/09/2000)
  3. Fabio Brugnara; Mauro Cettolo; Marcello Federico; Diego Giuliani,
    A Baseline for the Transcription of Italian Broadcast News,
    ICASSP 2000,
    , pp. 1667-
    , (ICASSP 2000,
    Istanbul, Turkey,
    05/06/2000 - 09/06/2000)
  4. Mauro Cettolo,
    You talk, I translate! Technical Aspects of the ITC-irst Speech Translation System,
    What is the state of the art in the speech translation field? Can a tool simultaneously translate what you are saying in your own language into any other language? And if so, could it be powerful enough to be put on the market? Undoubtly, this project is rather ambitious, almost on the boundary between `Science` and `Science-Fiction`. Nevertheless, I would like to describe to some degree of technicality, the ITC-irst speech translation system which has been demonstrated a number of times during 1999. What I hope is, primarily, to give the reader a summary of the current state in this field of research and, possibly, to try to locate the boundary between `fiction` and `reality`,
  5. Mauro Cettolo; A. Corazza,
    History Integration Into Semantic Classification,
    Computational Models of Speech Pattern Processing,
    , pp. 356 -
  6. Mauro Cettolo; A. Corazza; Gianni Lazzari; Fabio Pianesi; Emanuele Pianta; L. Tovena,
    A Speech-to-Speech Translation based Interface for Tourism,
    ENTER `99,
    , pp. 191-
    , (ENTER `99,
    Innsbruck, Austria,
    20/01/1999 - 22/01/1999)
  7. J. Haas; V. Warnke; H. Niemann; Mauro Cettolo; A. Corazza; Giuseppe Falavigna; Gianni Lazzari,
    Semantic Boundaries in Multiple Languages,
    Eurospeech `99,
    , pp. 535-
    , (Eurospeech `99,
    Budapest, Hungary,
    05/09/1999 - 09/09/1999)
  8. A. Corazza; Mauro Cettolo; Gianni Lazzari; Emanuele Pianta; Fabio Pianesi; Tovena L.M.,
    The ITC-irst Speech Translation System,
    Workshop AI*IA - Elaborazione del Linguaggio e Riconoscimento del Parlato `99,
    , pp. 30-
    , (Workshop AI*IA - Elaborazione del Linguaggio e Riconoscimento del Parlato `99,
    Trento, Italy,
    16/12/1999 - 17/12/1999)
  9. Mauro Cettolo,
    Segmentation and Classification of Italian Audio Broadcast News,
    This work presents first results in segmenting and classifying an Italian audio broadcast news corpus, under development at ITC-irst. The approach is based on two stages: during the first one, a HMM decoder segments and classifies the input audio stream in terms of acoustic source (music, speech, speech in presence of music). The second stage, based on the BIC method, detects speaker changes inside individual speech segments obtained through the first stage. On the test set, frame classification accuracy is 90.6%, while recall and precision of spectral changes detection with a tolerance of +-300cs are 82.9% and 65.6%, respectively,
  10. Mauro Cettolo; A. Corazza; R. De Mori,
    Language Portability of a Speech Understanding System,
    vol. 12,
    n. 1,
    , pp. 1 -